Thread by @benimmo

Thread: Must-read piece on state-sponsored trolling from @business.

It's coordinated. It's cross-platform. It's aimed at silencing critical or analytical voices. It's worldwide.

And the real work on countering it is just beginning.

bloomberg.com/features/2018-…

Here's the full report, thanks to @iftf and the great @carlynyst, @TheNickMon, @samuelwoolley and @camillefrancois.

There aren't many teams more world-class than this.

iftf.org/statesponsored…

States use virtual hate mobs, surveillance, misinformation, anonymous threats, and the invasion of victims’ privacy.

The ultimate goal? Kill free speech on the platforms, by using their own free-speech rules against them.

A small thing, but so important: defining the terms.

A "troll" isn't just someone who disagrees. It's one who deliberately targets the victim with hate and harassment.

And the crucial difference between intended and unintended falsehood, which is the toughest thing to prove.

Put this another way: repressive states have turned online platforms into a battleground.

There's a reason people talk of "troll armies" and "legions".

This tabulation of the attacks is spot on. (Full doc, pp. 12-13.)

- Target critics and troll analysts;
- Accuse of collusion with foreign intelligence;
- Accuse of treason;
- Rape and death threats;
- Memes.

China, Russia and Turkey, "dark paragons of disinformation" on the state level.

And more and more states and groups are queueing up to join.

Targeted harassment is relatively easy to spot from the traffic surges. Finding who initiated it is the bigger challenge.

The good thing is, it can be done. Not all trolls are smart.

medium.com/dfrlab/trolltr…

The challenge gets bigger when it's organised on a different platform.

Big question for the main platforms: how to respond when the harassment is orchestrated somewhere else, but launched on their site?

medium.com/dfrlab/trolltr…

Responses have to start by making it harder for trolls and bot herders to operate.

There are obvious abuses out there, like partisan political posts in Mexico getting thousands of likes from Brazil.

medium.com/dfrlab/electio…

The good news: the environment is already less permissive than two years ago.

Look how this botnet tried to escape detection. Quoting Jane Austen at random, while advertising porn.

It's not a major extra cost to the bot herder, but it's incremental.

medium.com/dfrlab/botspot…

But to start getting a handle on this, we need to see:

- more researchers;
- faster identification and analysis of troll spikes;
- better methodology for attribution;
- better victim support.

Above all, responses need to be cross-platform, the way the attacks are.

The dark paragons of disinformation cooperate to abuse the platforms. The platforms and researchers need to cooperate more to stem the abuse. /

Thread ends.